Freya

Freya develops voice AI agents that handle inbound and outbound telephone conversations for enterprises, operating 24/7 and supporting dozens of languages and local dialects. The platform covers the full agent lifecycle — training, fine-tuning, testing and deployment — and offers configurations such as an outbound sales agent that contacts prospects, handles objections and logs interactions, and an inbound welcome agent that answers queries, follows defined scenarios and updates contact records.

Typical deployments described by the company include debt collection and payment reminders, policy, coverage, claims and application status inquiries, campaign information and FAQ handling, appointment scheduling, rescheduling and reminders, lead qualification, renewals, quote intake and claim or loan servicing. Freya connects to existing CRM, telecommunications and IVR systems, as well as ticketing platforms, knowledge bases, policy databases, AMS and payment systems, retrieving live data such as account or policy details, updating records and triggering follow-up workflows. Compliance features cited include ingestion of firm-specific policy manuals, regulatory guidelines and operational procedures, call logging and auditability, end-to-end encryption, role-based access control, audit logging, voice verification and SMS two-factor authentication, with references to KVKK, GDPR and SOC compliance. Deployments are offered in both cloud and on-premise form.

Founding story

Co-founders Tunga Bayrak and Tomáš Nepala were roommates at the University of Pennsylvania and started Freya as their second company together. Bayrak, the CEO, previously worked as an AI research engineer building vision-language models for geometric reasoning and on audio transformers, wrote a paper on quantum machine learning at 16, and was a teaching assistant and AI research team member at Penn before dropping out. Nepala, the COO, studied at Wharton and worked across financial services firms including the insurance broker Gallagher. Both cited their exposure to customer support teams struggling with high call volumes as the motivation for the company.

Business model

B2B software: enterprises deploy Freya's voice AI agents in place of, or alongside, traditional call center staffing, integrating the agents into existing CRM, telephony and back-office systems.

Traction

Reported team size of 14 and Y Combinator Summer 2025 batch participation, followed by a $3.5 million seed round. Customer testimonials on the company site are attributed to a chief executive officer and to a managing director of an embedded BNPL product that uses Freya for outbound reminders and inbound FAQ handling without a call center. The company cites figures of over 50% lower cost than traditional call centers, 99% correct answers and a zero word error rate in comparison tests.

Latest developments

In November 2025 Freya announced a $3.5 million seed round with participation from Y Combinator, Rebel Fund, 212 VC, N1 Tech, BD Partners and angel investors, and was hiring for a DevOps engineer role in Istanbul.